Who? - Collaborative Web Search: Who, What, Where, When, and Why - page 13

example, Amershi and Morris ( 2008 , 2009 ) have found evidence that populations such as the elderly,
recent immigrants, and people in developing countries may engage in collaborative search activities,
but very little detailed data is currently available on the search needs and practices of these groups.
Specific sub-populations of information workers, such as legal professionals ( Hansen and Jarvelin ,
2005 ), may also have specialized needs for search tools, and they merit further study. More survey
and interview studies to learn about users would be extremely valuable, as much of the current data
is drawn from the student and information-worker populations, whose demographics and needs
may differ from those of other groups. Gathering data on collaborative tendencies for more detailed
demographic factors, such as by gender, age, geography, occupation, education level, or technical
expertise, may also facilitate new insights.
In terms of group configuration, few systems take advantage of nuanced relationships among
group members, either assuming either peer-based relationships or simple “guided search” relation-
ships. Systems like Cerchiamo ( Pickens et al. , 2008 ) with its “prospector” and “miner” roles, or like
Smart Splitting ( Morris et al. , 2008 ), which considers the expertise areas of each group member, are
beginning to uncover the potential for role-tailored group search systems, but this is still a rich and
valuable area for further investigation. Reflecting on roles within a collaborative search is an exciting
opportunity to design and study systems that support a variety of relationships among group mem-
bers, such as parents working with children or teachers with their students, as well as in potentially
inventing new roles for dividing labor among group members that do not currently exist.
Additionally, although some work has been done on understanding current trends regarding
group size, there is room for deeper understanding of not only what size groups users form, but
also what size groups are optimal for different tasks, and how group size is being influenced or
constrained by current technologies. Understanding how the performance of proposed collaborative
search systems scales as group size changes is another open issue.